market-research
This skill should be used when the user asks for market research, market sizing, TAM/SAM/SOM, market trends, industry analysis, or market segmentation — e.g. "how big is this market", "do market resea
SKILL.mdventure/skills/market-research/SKILL.md
GitHubMarket Research — Sizing, Trends & Segments
Produce decision-grade market research: how big is the opportunity, how is it changing, which segment to enter first. Numbers must be traceable (cited) and honest (confidence-rated) — a wrong TAM quietly poisons every later module.
Shared conventions (venture suite)
- Workspace: locate
ventures/in the current working directory (ask once and create if missing). All outputs go toventures/<slug>/<NN-module>/. - Templates are read-only. Files under a skill's
references/live inside the installed plugin and are shared by every venture. Never write to them. "Fill template X" always means: copy it intoventures/<slug>/<NN-module>/under a dated name and fill the copy. - Read
VENTURE.mdfirst. When done: update its Status table and append key findings to "Key facts & numbers" (format:- [module, YYYY-MM-DD] finding (source, confidence)). - Language: converse in the user's language. Write formal deliverables (reports, canvases, battlecards) in the
report_languageset in VENTURE.md. Persian reports keep standard English business terms as-is (TAM, SAM, SOM, CAC, LTV, GTM, churn, retention...). - Research integrity: cite every external claim (source + date). Mark every estimate with confidence (high/med/low). Never fabricate numbers — write "no data found" instead.
- Chain: end by updating VENTURE.md and suggesting the logical next module.
Workflow
1. Scope (with the user, briefly)
From VENTURE.md + one quick exchange, pin down:
- Market definition — the job being done, not the product category ("appointment no-show reduction for clinics", not "SaaS").
- Geography — local market vs global changes sources and method entirely.
- Mode —
quick(~10 min scan: top-line size, 3–5 trends, segment list) ordeep(full report). Default: quick forideastage, deep otherwise. State which mode you're running.
2. Research
Consult references/data-sources.md for where to look and references/sizing-guide.md for method. Core rules:
- Local markets → local language. For Iran search in Persian (آمار, گزارش بازار, مرکز آمار ایران, cafebazaar/digikala data); for Turkey in Turkish; etc. English-only search badly undercounts local markets.
- Triangulate: any number that matters needs ≥2 independent sources, or gets
confidence: lowexplicitly. - Recency: prefer data ≤2 years old; note the year next to every figure.
Quick mode: run 5–10 targeted WebSearch/WebFetch rounds inline yourself.
Deep mode: decompose into 3–5 research questions (e.g. ① market size & growth, ② segments & buyer behavior, ③ trends/drivers/regulation, ④ analogous markets & benchmarks) and fan out venture:market-researcher agents in parallel — one question each, single message, multiple Agent calls. In each agent prompt include: the venture one-liner, market definition, geography, the specific question, and expected output format (findings + citations + confidence). Synthesize results yourself afterward.
3. Size the market — both directions, show the math
Follow references/sizing-guide.md:
- Top-down: industry reports → TAM → filter to SAM → realistic share → SOM.
- Bottom-up (more credible, always include): (number of target customers) × (reachable %) × (realistic price). Build each factor from cited data.
- The two methods will disagree — explain why, present a range, and state which number you'd bet on.
- Every formula appears in the report with its inputs, so the user (or their client) can re-derive it.
4. Write the report
Fill a copy of references/report-template.md, saved as ventures/<slug>/02-market/market-research-YYYY-MM-DD.md (quick mode → market-scan-YYYY-MM-DD.md, abbreviated sections). Write in report_language. Re-runs create a new dated file — never overwrite history.
5. Close the loop
- Update VENTURE.md: Status row →
done YYYY-MM-DD; append ≤10 key facts (SOM range, growth rate, top trend, chosen entry segment...); add new risky assumptions discovered. - Suggest next: usually
/venture:competitors(the sizing surfaced who's competing) or/venture:customersif buyer behavior data was thin.
Quality bar
- A skeptical investor should find zero unsourced numbers.
- "No reliable data found — proxy used: X (low confidence)" is a good sentence; an invented number is a firing offense.
- Recommend an entry segment explicitly — research that ends without a "so what" is decoration.